Spokesperson Effects in High Involvement Markets

Citation:

Timothy B. Heath, David L. Mothersbaugh, and Michael S. McCarthy (1993) ,"Spokesperson Effects in High Involvement Markets", in NA - Advances in Consumer Research Volume 20, eds. Leigh McAlister and Michael L. Rothschild, Provo, UT : Association for Consumer Research, Pages: 704-708.

Advances in Consumer Research Volume 20, 1993      Pages 704-708

SPOKESPERSON EFFECTS IN HIGH INVOLVEMENT MARKETS

Timothy B. Heath, University of Pittsburgh

David L. Mothersbaugh, University of Pittsburgh

Michael S. McCarthy, University of Pittsburgh

[This research was partially supported by a research grant from the Katz Graduate School of Business.]

Existing theory and research suggest that persuasive cues such as spokesperson fame are ineffective when consumers engage in issue-relevant thinking. However, these studies examined behavior in noncompetitive, between-subjects settings. Experiment 1 replicates the procedures of past research, whereas Experiment 2 mimics marketplace competition by varying spokesperson fame within subjects. Spokesperson fame had significant effects on brand attitudes and choice only in Experiment 2's competitive setting. The data suggest that (1) existing theory be extended with the concept of cue neutralization, (2) attitude research use within-subjects designs to capture marketplace persuasion processes, and (3) persuasive cues such as spokesperson fame are increasingly important as markets mature and competitive parity increases.

Spokesperson effects have been studied by social psychologists and consumer researchers (Hovland and Weiss 1951; Kahle and Homer 1985). Existing research shows that spokespeople influence consumers primarily when consumers are unable to base their evaluations on issue-relevant thinking (Batra and Ray 1986; Chaiken 1980; Petty and Cacioppo 1981, 1986; Greenwald and Leavitt 1984). However, practitioners often use spokespeople to promote products, even those products commanding considerable pre-purchase deliberation such as automobiles. The current study proposes and tests competitive effects that partially explain this discrepancy.

Unlike the environments studied in traditional attitude research, those faced by practitioners are often competitive. Consumers see multiple ads and brands within each product class. Whether competition influences the effectiveness of spokesperson fame is investigated here. As in past research, Experiment 1 tests spokesperson effects in noncompetitive settings. Experiment 2 then tests the moderating role of competition by replicating Experiment 1's stimuli and procedures in a competitive setting.

THEORY AND RESEARCH

Traditional attitude research has documented many variables affecting attitude change including personality (Hovland and Janis 1959), classical conditioning (Staats and Staats 1958), cognitive dissonance (Festinger 1957), self perception (Bem 1972), and cognitive responses (Greenwald 1968). Many of these effects have been subsumed under broader theories recognizing the wide array of processes through which attitudes are formed and changed.

Primary among the multi-process theories are Chaiken's (1980) Systematic/Heuristic Model and Petty and Cacioppo's (1981, 1986) Elaboration Likelihood Model (see also Greenwald and Leavitt 1984). These models distinguish between attitude change based on product features and attitude change based on less substantive stimuli such as spokesperson fame. The processes underlying these two are referred to respectively as systematic and heuristic processes by Chaiken, and as central and peripheral routes to persuasion by Petty and Cacioppo.

Multi-process theories of attitude change maintain that peripheral cues are relatively ineffective when consumers engage in issue-relevant thinking. This is consistent with earlier research on communication effects (Krugman 1965; Ray et al. 1973; Robertson 1976) and has received considerable empirical support (for reviews see Greenwald and Leavitt 1984; Petty and Cacioppo 1986). For example, famous spokespeople have been found to improve brand attitudes in low but not high involvement (e.g., Petty et al. 1983).

Evidence of limited peripheral-cue effects in high involvement is confined primarily to experimental designs where peripheral cues are manipulated between subjects. Subjects typically see one ad for one brand within a given product class, where the ad contains either a weak or a strong peripheral-cue (e.g., nonfamous or famous spokesperson). Each subject then evaluates that brand (e.g., Petty et al. 1983). In contrast, however, peripheral cues vary within subjects in the marketplace. Each consumer sees multiple brands with corresponding ads and cues. In such competitive situations, a persuasive ad is of little use if competitors' ads are even more persuasive. In the marketplace, therefore, persuasion relative to that of the competition is more important than absolute persuasion.

Choice research inherently addresses relative persuasion. Regardless of whether the process is compensatory or noncompensatory, choice implies cross-alternative comparisons. Whereas absolute persuasion is effected little by the distribution of attribute levels across alternatives, relative persuasion is. The less variation in attributes across alternatives, the less influence each has on relative persuasion, especially when measured with choice (Tversky 1972). The tendency for consumers to ignore attributes held constant across alternatives has been referred to as cancellation in the study of risky decisions (Kahneman and Tversky 1979).

Similar processes may exist in multi-attribute product contexts. We propose that cue neutralization often occurs in the marketplace. For example, if three television brands have identical levels of picture resolution and warranty, the persuasiveness of Brand A's warranty is neutralized by the warranties of the other brands. If all central cues are neutralized, then peripheral cues may play a critical role in the decision since there are no other features on which the brands can be differentiated. As central cues become increasingly neutralized, peripheral cues are hypothesized to become more important ceteris paribus, regardless of the level of issue-relevant thinking.

Homogenous product features reflect the simplest type of cue neutralization; namely within-attribute cue neutralization: Television A's picture resolution is neutralized by Television B's picture resolution. However, cross-attribute cue neutralization is also possible, and distinguishes neutralization from Kahneman and Tversky's (1979) cancellation. For example, Television A might have superior picture resolution, while Television B has superior sound quality. If the televisions are liked comparably, then Television B's superior sound quality is said to neutralize the persuasiveness of Television A's superior picture resolution. Consumer preferences across brands are homogenous despite product heterogeneity (cf. Slovic 1975). As with within-attribute cue neutralization, as cross-attribute neutralization of central cues increases, the power of peripheral cues is expected to increase regardless of the level of issue-relevant thinking.

The present study examines the effects of spokespeople in between-subjects (Experiment 1) and within-subject environments (Experiment 2). Experiment 2 assesses spokesperson effects in the face of within-attribute cue neutralization. This is important for both theory and practice. For theory, peripheral-cue effects in the face of issue-relevant thinking would suggest extending multi-process theories with the concept of cue neutralization. For practice, existing evidence suggests that U.S. markets are increasingly characterized by competitive parity and homogenous products (Wall Street Journal, October 19, 1989, Page B1). If spokesperson effects occur in such settings, peripheral cues may become more important in the marketplace as markets mature regardless of involvement.

STUDY OVERVIEW AND HYPOTHESES

Two experiments assess the power of spokespeople in a high-involvement competitive environment where central cues are neutralized within attributes. These experiments test spokesperson effects on brand attitudes using both between-subjects (Experiment 1) and within-subjects manipulations (Experiment 2). The within-subjects manipulation of Experiment 2 further assesses the effects of spokesperson fame on brand choice.

The between-subjects manipulation of Experiment 1 replicates the procedures of traditional attitude research reporting a lack peripheral-cue effects in high involvement. It therefore examines the following null hypothesis:

H1: Spokesperson fame will not influence attitudes when manipulated between subjects and when there is issue-relevant thinking.

Experiment 2 uses a within-subjects replication of Experiment 1 to mimic marketplace competition. Between-subjects manipulations have been the primary source of empirical support for existing multi-process attitude theories. Whether the predictions of these theories hold in environments more akin to the marketplace remains to be tested. Therefore, Experiment 2 tests the following hypothesis:

H2: Spokesperson fame will influence brand attitudes when manipulated within subjects as long as central cues are neutralized, even when there is issue-relevant thinking.

Experiment 2 further tests the effects of spokespeople on choice:

H3: Spokesperson fame will influence choice probabilities as long as central cues are neutralized, even when there is issue-relevant thinking.

EXPERIMENT 1

Method

Pretests. Twenty-five student subjects at a large eastern university rated the fame of six potential spokespeople on a 9-point ratings scale ranging from 1 (not famous) to 9 (famous). Two of the spokespeople were considered famous a priori (Michael J. Fox and Jay Leno), whereas the others were fictitious (Alex Tyler, Darrell Spencer, Tom Hite, and John Meyers). As expected, Michael J. Fox and Jay Leno were rated as more famous than the others. Therefore, Michael J. Fox and Alex Tyler were selected for use in the experiments (M's = 7.6 and 2.7, respectively; t₍₂₄₎ = 10.97). [All inferential statistics are evaluated with respect to a=.05.]

Subjects and Design. Experiment 1 varied spokesperson fame (nonfamous vs. famous) between subjects. Twenty-one subjects enrolled at a large eastern university were randomly assigned to each of the two conditions. When spokesperson fame was manipulated, two similar versions of ad copy were counterbalanced. The counterbalancing of similar ad copies was needed to reduce demand in Experiment 2's within-subject's manipulation, and was included in Experiment 1 to make the two experiments comparable. The second version of the ad copy for the spokesperson conditions was created by making minor wording changes to the first.

Procedure. The first section of the booklet consisted of an introductory cover story to provide a rationale for the study and heighten involvement. Students were told that cruises were becoming increasingly popular, and that their responses would be instrumental in the development of the products or services they evaluated.

Following the cover story, subjects read an advertisement for the brand they were to evaluate. The ad was in the form of a radio script. Using radio scripts helped eliminate confounding effects of elements typically found in print ads (photos, layout, typeface, etc.) and broadcast ads (music, editing, camera angles, etc.). The ad used a disguised brand name (Cruise L) to avoid the effects of brand name familiarity or liking on product evaluations. The scripts identified the announcer (spokesperson) as Michael J. Fox or Alex Tyler.

After reading the radio script, subjects examined a table containing attribute information about the cruises. The information addressed eight cruise attributes (price, overall quality, number of days, number of islands, night club, Olympic pool, athletic activities, casino). As in prior research, attribute information was held constant across nonfamous and nonfamous spokesperson conditions.

After reviewing the ad and the attribute table, subjects evaluated the cruise on 9-point bad/good, unsatisfactory/satisfactory, and unfavorable/favorable scales ranging from -4 to +4 (cf. Petty et al. 1983). They then responded to a check on the manipulation of spokesperson fame, a measure of issue-relevant thinking discussed later, and experimental demand.

Results

Demand Check. The demand check asked subjects if they felt there was an alternative purpose to the study. If they answered yes, they were then asked to describe the alternative purpose. None of the subjects mentioned an alternative purpose that was remotely related to the manipulation of spokesperson fame.

Manipulation Checks. The manipulation of spokesperson fame was checked with the same 9-point scale used in the pretest. As in the pretest, Michael J. Fox was rated as more famous than Alex Tyler (M's = 7.6 and 2.4, respectively; t₍₁₉₎ = 8.2). [Throughout the analyses, the data are collapsed across the counterbalanced variables. There were no main effects of the counterbalanced variables, and interactions with them (1) did not alter the interpretations of spokesperson effects, (2) were small, and (3) were uninterpretable.] Thus, the manipulation of spokesperson fame was successful.

To test the study's hypotheses, it was critical that consumers engage in issue-relevant thinking. Issue-relevant thinking was checked with two items. The first measured how carefully subjects evaluated the brand compared to the care used in evaluating household products such as paper towels, laundry detergents, and cleansers. The second measured how carefully subjects evaluated the brand compared to the care used in evaluating consumer electronics such as televisions, stereo equipment, and microwave ovens. These two sets of product classes were considered representative of relatively low and high involvement products, respectively.

The two thinking scales ranged from -8 (much less carefully) to 0 (about as carefully) to +8 (much more carefully), and indicated that subjects engaged in issue-relevant thinking. Subjects evaluated cruises more carefully than household products, although the difference did not achieve statistical significance (M = 1.0; t₍₂₀₎ = 1.3). Likewise, subjects evaluated cruises less carefully than consumer electronics products, although again the difference was not statistically significant (M = -1.5; t₍₂₀₎ = -1.75).

Spokesperson Effects. Consistent with the predictions of multi-process theories, Hypothesis 1 predicted that spokespeople would not influence brand attitudes under high involvement. The three items used to measure brand attitudes were combined due to high reliability (Cronbach's a = .97). As expected, there was no effect of spokesperson fame on brand attitudes. Brand attitudes were virtually unaffected by peripheral cues varied between-subjects (M_Nonfamous = 2.6 and M_Famous = 2.5). [One concern is that the relatively small n's rendered insufficient statistical power. Two factors mitigate against this concern: (1) the small samples had sufficient power to verify the spokesperson manipulation, and (2) brand-attitude means were directionally inconsistent with the manipulation checks and Hypothesis 2.]

Discussion

Although impossible to prove a null hypothesis, Experiment 1 suggests that spokesperson fame does not affect attitudes when (1) it is manipulated between subjects, (2) product features are held constant across conditions, and (3) consumers engage in issue-relevant thinking. This is consistent with many prior between-subjects assessments reporting that peripheral cues are ineffective in high involvement (e.g., Petty et al. 1983).

Since brand attributes were easy to understand and since subjects were given unlimited time to examine them, subjects should have possessed sufficient ability and opportunity to process brand information. Given this, the ineffectiveness of spokesperson fame supports our contention that it is, in fact, a peripheral cue. Whether such a peripheral cue is ineffective in competitive environments is assessed in Experiment 2.

EXPERIMENT 2

Experiment 1 replicated past research and offered further support for the following principle of multi-process theories: Peripheral cues such as spokesperson fame are relatively ineffective in high involvement. However, Experiment 1 and prior studies supporting this principle do so in noncompetitive environments and thereby assess absolute persuasion. They do not allow for competitive processes such as cue neutralization to further moderate the effects of peripheral cues. To test such effects, Experiment 2 used the same peripheral and central cues as Experiment 1, held central cues constant as in Experiment 1, and varied spokesperson fame within subjects.

Experiment 2 tested Hypotheses 2 and 3. They predict that spokespeople will influence attitudes and choices in the face of considerable issue-relevant thinking if central cues are neutralized. Cue neutralization was affected in Experiment 2 by configuring two of the four competing brands to be equally liked and liked more than the others.

Method

Subjects and Design. Experiment 2 varied spokesperson fame (nonfamous vs. famous) within subjects. Subjects saw four brands, two of which had comparable product features (target brands) and were superior to the other two (nontarget brands). Two counterbalanced variables were manipulated between subjects: (1) task order (attitude then choice vs. choice then attitude), and (2) the target brand receiving the stronger peripheral cue (Brand L vs. Brand N). Fifty-nine students enrolled at a large eastern university were randomly assigned to the four conditions resulting from the counterbalancing.

Relative to a between-subjects manipulation, a within-subjects manipulation provides many benefits including increased statistical power and assessments of competitive effects. However, it also increases the transparency of the manipulation and thereby increases the potential for experimental demand. Two procedures were used to reduce demand artifacts (Shimp, Hyatt, and Snyder 1991). First, subjects indicating any sense of the experiment's intent were dropped. Second, and based on the pretests, nontarget brands and ads were configured to disguise the experiment's intent. Spokesperson fame varied across target brands (Michael J. Fox vs. Alex Tyler) and nontarget brands (Jay Leno vs. Tom Hite). Moreover, the copy varied across the four ads. Two similar versions were counterbalanced across the two target brands. Thus, subjects saw four ads that differed in copy and spokesperson. Two used a famous spokesperson while two used a nonfamous spokesperson.

Stimuli and Procedure

To compare the results from Experiment 1 with those from Experiment 2, the stimuli and procedures used in the two experiments had to be as similar as possible. Therefore, Experiment 2 used the spokespeople and ad copies of Experiment 1 to manipulate spokesperson fame and counterbalance ad copy. Experiment 2 also used the same radio-ad format and the same levels of product features for the two target brands.

Subjects first examined ads for each brand. In order to avoid any effects of brand name liking or familiarity, the brands were labeled Cruise J, L, N, and P. Immediately following the four ads was a table containing the attribute values for each of the four advertised products (central cues).

The tables were constructed subject to various constraints. The two target brands were superior to the two nontarget brands. Further, to facilitate cross-experiment comparisons, the product features of the target brands were identical to those used in Experiment 1. Thus, features did not vary across target brands. This increased the comparability of Experiment 2's within-subjects cue manipulation with the between-subjects cue manipulation of Experiment 1 and prior research supporting multi-process theories. However, unlike between-subjects designs where consumers never see comparable products, exposure to comparable products in Experiment 2 was expected to neutralize central-cue effects and thereby increase the power of peripheral cues.

After reviewing the ads and brand information, subjects (1) chose the brand they would buy if they were in the market today and (2) reported their attitudes toward each of the four brands. To control for carryover from one measure to the next, task order was counterbalanced. Half of the subjects reported attitudes before making their choices, while the other half chose first. Brand attitude measures, the demand check, and manipulation checks were the same as those in Experiment 1.

Results

Demand Check. Nine of the fifty-nine subjects mentioned an alternative purpose of the study that was at least remotely akin to the manipulation of spokesperson fame. These subjects were dropped from subsequent analyses although this had no effect on the conclusions.

Manipulation Checks. The manipulation of spokesperson fame was checked with the same 9-point scale used in the pretests and Experiment 1. Michael J. Fox was rated as more famous than Alex Tyler (M_Famous = 7.7 vs. M_Nonfamous = 2.0; t₍₄₈₎ = 16.2).

Issue-relevant thinking was checked with the same measures used in Experiment 1. Subjects evaluated the cruises more carefully than household products (M = 3.3; t₍₄₉₎ = 5.4) and about as carefully as high-involvement consumer electronics products (M_Cruise = -.48; t₍₄₉₎ = -.85). Thus, subjects did in fact engage in issue-relevant thinking.

Spokesperson Effects on Attitudes. Hypothesis 2 proposed that peripheral cues would influence brand attitudes in high involvement if central cues were neutralized. This was tested by analyzing a composite brand attitude measure consisting of the average of the three brand attitude items (Cronbach's a = .93). Consistent with Hypothesis 2, spokesperson fame increased brand attitudes (M_Nonfamous = 2.9 and M_Famous = 3.3; t₍₄₆₎ = 3.02).

Spokesperson Effects on Choice. In extending multi-process persuasion theories to a within-subjects environment with central-cue neutralization, we hypothesized that spokesperson fame would increase choice probabilities (Hypothesis 3). The effects of spokesperson fame on choice were assessed with a LOGIT model. Consistent with Hypothesis 3, spokesperson fame increased the likelihood of brand choice (c²(1) = 16.9). Brands with the famous spokesperson were chosen 82% of the time.

GENERAL DISCUSSION

Summary

Experiment 1 replicated prior null effects of spokesperson fame in noncompetitive, between-subjects settings. When there is no competition but issue-relevant thinking exists, peripheral cues have little effect on brand evaluations. Experiment 2 examined spokesperson fame in a competitive environment where fame was manipulated within subjects such that each consumer saw various levels of spokesperson fame across competing brands. As in Experiment 1, central cues were held constant across the two target brands which, in a competitive within-subjects environment, was expected to affect central-cue neutralization. As hypothesized, the same peripheral cues that were ineffective in Experiment 1 were, in fact, persuasive in Experiment 2. They improved brand attitudes and increased market share despite considerable issue-relevant thinking.

Evidence of peripheral-cue effects in high involvement markets is important for three reasons. First, the effects of competition require modifications to existing multi-process theories of attitude change. Popular principles regarding the persuasiveness of peripheral cues in high involvement must be tempered by the level of central cue neutralization in the marketplace. Second, the results suggest that if attitude research is to generalize to the marketplace, within-subjects manipulations are required. Third, the results suggest a growing importance of peripheral cues since U.S. markets evidence increasingly high levels of product homogeneity. As markets continue to mature and central cues become more neutralized, peripheral cues should become more powerful regardless of involvement.

Limitations and Future Research

The current study suffers from the use of a single peripheral cue and product class. The generalizability of the findings must therefore be expanded in future research by using a wider range of stimuli. Furthermore, the current study assessed spokesperson effects when central cues were neutralized through the use of homogenous brands. Future research must assess whether peripheral-cue effects exist in the face of issue-relevant thinking and heterogenous brands. Since negative correlations across attributes make decisions particularly difficult (e.g., Johnson, Meyer, and Ghose 1989), it is likely that such effects will prevail. Finally, future research is needed to assess other potential causes of Experiment 2's peripheral-cue effects. For example, within-subjects manipulations may introduce perceptual contrast and thereby increase the perceived difference between peripheral-cue levels (Lynch, Chakravarti, and Mitra 1991).

REFERENCES

Batra, Rajeev and Michael L. Ray (1986), "Situational Effects of Advertising Repetition: The Moderating Influence of Motivation, Ability, and Opportunity to Respond," Journal of Consumer Research, 12 (March), 432-445.

Bem, Daryl J. (1972), "Self-Perception Theory," in Advances in Experimental Social Psychology, Vol. 6, ed. Leonard Berkowitz, New York: Academic Press, 2-57.

Chaiken, Shelly (1980), "Heuristic Versus Systematic Information Processing and the Use of Source Versus Message Cues in Persuasion," Journal of Personality and Social Psychology, 42 (January), 116-131.

Festinger, L. (1957), A Theory of Cognitive Dissonance, Stanford, CA: Stanford University Press.

Greenwald, Anthony G. (1968), "Cognitive Learning, Cognitive Response to Persuasion, and Attitude Change," in Psychological Foundations of Attitudes, eds. Anthony G. Greenwald, Timothy C. Brock, and Thomas Ostrom, New York: Academic Press, 147-170.

Greenwald, Anthony G. and Clark Leavitt (1984), "Audience Involvement in Advertising: Four Levels," Journal of Consumer Research, 11 (June), 581-592.

Hovland, Carl I. and Irvin L. Janis, eds. (1959), Personality and Persuasability, New Haven: Yale University Press.

Hovland, Carl I. and W. Weiss (1951), "The Influence of Source Credibility on Communication Effectiveness," Public Opinion Quarterly, 15, 635-650.

Johnson, Eric J., Robert J. Meyer, and Sanjoy Ghose (1989), "When Choice Models Fail: Compensatory Models in Negatively Correlated Environments," Journal of Marketing Research, 26 (August), 255-270.

Kahle, Lynn R. and Pamela M. Homer (1985), "Physical Attractiveness of the Celebrity Endorsers: A Social Adaptation Perspective," Journal of Consumer Research, 11 (March), 954-961.

Kahneman, Daniel and Amos Tversky (1979), "Prospect Theory: An Analysis of Decision Under Risk," Econometrica, 47 (March), 263-291.

Krugman, Herbert E. (1965), "The Impact of Television Advertising: Learning without Involvement," Public Opinion Quarterly, 29 (Fall), 349-356.

Lynch, John G. Jr., Dipankar Chakravarti, and Anusree Mitra (1991), "Contrast Effects in Consumer Judgments: Changes in Representations or in the Anchoring of Rating Scales," Journal of Consumer Research, 18 (December), 284-297.

Petty, Richard E. and John T. Cacioppo (1981), Attitudes and Persuasion: Classic and Contemporary Approaches, Dubuque, IA: William C. Brown.

Petty, Richard E. and John T. Cacioppo (1986), Communication and Persuasion: Central and Peripheral Routes to Attitude Change, New York, Springer.

Petty, Richard E., John T. Cacioppo, and David Schumann (1983), "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement," Journal of Consumer Research, 10 (September), 135-146.

Ray, Michael L., Alan G. Sawyer, Michael L. Rothschild, Roger M. Heeler, Edward C. Strong, and Jerome B. Reed (1973), "Marketing Communication and the Hierarchy of Effects," in New Models for Mass Communication Research, Vol. 2, ed. Peter Clarke, Beverly Hills, CA: Sage Publications, 147-176.

Robertson, Thomas S. (1976), "Low-Commitment Consumer Behavior," Journal of Advertising Research, 16 (April), 19-24.

Shimp, Terence A., Eva M. Hyatt, and David J. Snyder (1991), "A Critical Appraisal of Demand Artifacts in Consumer Research," Journal of Consumer Research, 18 (December), 273-283.

Slovic, Paul (1975), "Choice between Equally Valued Alternatives," Journal of Experimental Psychology: Human Perception and Performance, 1, 280-287.

Staats, A. W. and C. K. Staats (1958), "Attitudes Established by Classical Conditioning," Journal of Abnormal and Social Psychology, 57, 37-40.

Tversky, Amos (1972), "Elimination by Aspects: A Theory of Choice," Psychological Review, 79, 281-299.

----------------------------------------